Error: object 'previous.upper' not found

[Kuldeep Roy]

Hi,

I am trying to use GoodTurningMD to evaluate the convergence of my MD. I tried to run using the RMSD matrix generated using GRCARMA program. However, it shows the following verbose output and finally ends with errors:

1. Opening file for reading ...                 OK
2. Checking dimensions ...                      OK
3. Loading the matrix ...             5000 x  5000
4. Sanity checks :
   Non-zero matrix ?                           yes
   Negative values present ?                  none
   Sane RMSDs ?                                yes
   Origin of matrix at expected position ?     yes
   Average deviation from symmetry .......... pass
   Symmetry outliers ? ...................... none
Error: object 'previous.upper' not found

Could you please let me know where is the source of such error?

Thanks in advance,

Kuldeep

[Panagiotis Koukos]

Hello Kuldeep,

unfortunately the error message that is reported by R is generic enough that it is nearly useless. It usually means that your are trying to access something that doesn't exist, say a column of a dataframe by a name that is wrong or an operation similar to that. In this particular case it fails to read a value that it should be able to read from the RMSD matrix. This leads me to believe that there is something wrong with your input file or your version of R is different enough to all of the ones that I have tested on that it results in unpredictable behaviour.

I would suggest you send me the matrix file and I have a look at it in order to try and figure out what's wrong. If you do decide to send me the file, also make sure to include some information about your system as well (OS, hardware info, R version) and to compress the file too. Even with standard compression algorithms the zipped matrix shouldn't be bigger than ~50MB. Ideally upload it to Dropbox/Google Drive/other cloud equivalent rather than sending it as an attachment. If this is not an option due to low bandwidth, create the smaller matrix that still creates the error, and compress and send that one.

In case you decide to debug this on your own, I just tried the same analysis with some of our files and it worked without any errors. The machine that the analysis was performed on runs 64 bit linux with the latest version of grcarma/carma and R version 3.0.2. The steps I took were:

launching grcarma

running an RMSD analysis which creates an RMSD matrix with dimensions of 5000x5000

running the good-turing procedure on that matrix.

Just to make sure that this was not a fluke, try repeating the operation to make sure that it fails consistently. If it does, try creating matrices of different sizes and test it with those as well. If it is still failing, you will have to manually launch the R interface and switch to the carma_results_* directory that contains the results of your grcarma run. In that directory there will be a file named Good_Turing.R. That is the R script that performs the Good Turing analysis. You can launch it by typing "source("Good_Turing.R")" - without the external quotes. After that the program will ask you for the RMSD matrix file. Enter the path to the file without any quotes. If that was specified successfully, the analysis should have started. If it fails at the same point, make note of any errors that are returned. They might be more specific than the one that was returned in the grcarma text window.

Also try to update any relevant software, ie try to get a later version of R, although I have tested it with versions going back to 2.9 without any problems.

Something else that comes to mind is that if you are running this on a particularly old machine you might be having memory issues as for me the memory consumption during the processing of a 5Kx5K matrix peaks at 700MB.

All the best,

Panagiotis Koukos

--
You received this message because you are subscribed to the Google Groups "Good-Turing-MD" group.
To unsubscribe from this group and stop receiving emails from it, send an email to good-turing-m...@googlegroups.com.
To post to this group, send email to good-tu...@googlegroups.com.
Visit this group at http://groups.google.com/group/good-turing-md.
For more options, visit https://groups.google.com/d/optout.

[Kuldeep Roy]

Dear Panagiotis,
Thanks very much for your quick response. I tried running using
another RMSD.matrix file where it ran successfully.

Thanks again,

Kuldeep

--
Kuldeep K. Roy, Ph.D.
Post-doctoral Research Associate
Department of BioMolecular Sciences
School of Pharmacy
The University of Mississippi
436 Faser Hall, P. O. Box 1848
University, MS 38677, United States (US)
Email: kk...@olemiss.edu; Phone: +1-662-801-3850